NVIDIA MQM9700-NS2F 64-Port 400Gb/s InfiniBand Switch with SHARPv3 Technology in 1U Form Factor

Engineered for the most demanding high-performance computing (HPC) and artificial intelligence (AI) workloads, the MQM9700-NS2F is a fixed-configuration, 64-port 400Gb/s InfiniBand switch delivering unprecedented throughput and scalability in a 1U form factor. This network switch is a cornerstone for building extreme-scale, low-latency data center fabrics, accelerating research and innovation by moving computation into the network itself.

• Unmatched Density & Performance: 64 non-blocking 400Gb/s ports (aggregate 51.2 Tb/s bidirectional throughput).
• In-Network Computing: Features 3rd-gen NVIDIA SHARPv3 technology, accelerating collective operations by up to 32X.
• Ultra-Low Latency: Designed for real-time data exchange in highly parallelized algorithms.
• Topology Flexibility: Supports Fat Tree, DragonFly+, SlimFly, and multi-dimensional Torus topologies.
• High Availability: Hot-swappable, redundant power supplies and fan units for continuous operation.
• Advanced Manageability: On-board subnet manager for out-of-the-box deployment of up to 2,000 nodes.

The MQM9700 series leverages the full NVIDIA Quantum-2 platform stack. It is built upon NDR 400G InfiniBand technology, integrating advanced in-network computing via SHARPv3 (Scalable Hierarchical Aggregation and Reduction Protocol). Key supported technologies include:

• Remote Direct Memory Access (RDMA) for CPU bypass and ultra-efficient data movement.
• Adaptive Routing and congestion control for optimized network utilization.
• Port-Split Technology enabling 128 ports of 200Gb/s, offering cost-effective topology design.
• MLNX-OS for comprehensive switch management via CLI, WebUI, SNMP, and JSON-RPC.

This high-performance network switch is ideal for:

• AI/ML Training Clusters: Facilitating fast GPU-to-GPU communication for large-scale model training.
• High-Performance Computing (HPC): Accelerating scientific simulations, genomic research, and financial modeling.
• Hyperscale Data Centers: As a high-density Top-of-Rack (ToR) or spine switch in scalable fabrics.
• Cloud Infrastructure: Building high-throughput, low-latency private or public cloud backbones.
• High-Frequency Trading & Real-Time Analytics: Where microsecond latencies are critical.

• Industry-Leading Performance: Highest port density and throughput per 1U, reducing data center footprint and cost per gigabit.
• SHARPv3 In-Network Compute: Dramatically reduces data movement, accelerating AI and HPC applications directly within the network switch fabric.
• Future-Proof Design: Backward compatible with previous InfiniBand generations and supports flexible port splitting for 200Gb/s deployment.
• Resilient & Smart Fabric: Built-in self-healing, adaptive routing, and advanced congestion control ensure maximum application performance.
• Comprehensive Ecosystem: Supported by NVIDIA's UFM fabric management platform for end-to-end visibility and control.

We stand behind every MQM9700 unit we supply. Our commitment includes:

• Warranty: Standard 1-year manufacturer's warranty. Extended warranties available.
• Technical Support: 24/7 customer service and technical consultation via phone and email.
• Fast Delivery: Leverage our $10 million inventory for prompt shipment of the MQM9700-NS2F and related products like network card and cables.
• Integration Services: Our expert technical team can assist with network design, deployment, and optimization.

Q: What is the difference between the MQM9700-NS2F and MQM9790-NS2F models?
A: The MQM9700-NS2F is an internally managed switch with an on-board subnet manager, ideal for standalone or small cluster deployment. The MQM9790 is unmanaged and designed to be controlled by an external fabric manager like NVIDIA UFM for large-scale data centers.

Q: Can this switch connect to existing 200Gb/s or 100Gb/s InfiniBand equipment?
A: Yes. The MQM9700 series is backward compatible. Using appropriate cables or adapters, it can connect to NDR, HDR, EDR, and FDR generation network card and switches.

Q: Does it support Ethernet protocols?
A: No. This is a native InfiniBand switch. For converged Ethernet/InfiniBand solutions, please inquire about our NVIDIA Spectrum-X offerings.

Q: What type of cables are required?
A: It supports OSFP-based Direct Attach Copper (DAC), Active Optical Cables (AOC), and standard optical transceivers. Our team can help you select the correct cabling for your distance and budget requirements.

• Compatibility: Ensure your host servers are equipped with compatible NDR 400G or HDR/EDR InfiniBand network card (NIC card).
• Environmental: Adhere to specified operating temperature (0°C-40°C) and humidity (10%-85% non-condensing) ranges for optimal performance and longevity.
• Airflow & Rack Planning: Confirm your data center's cooling scheme matches the switch's airflow direction (P2C for this model). Ensure adequate clearance for hot-swappable components.
• Safety & Compliance: The product is certified for CE, FCC, RoHS, and cTUVus standards. Installation should be performed by qualified personnel.
• Software/Firmware: For the latest features and security, plan to update to the recommended MLNX-OS version upon deployment.

With over a decade of industry experience, we have established ourselves as a leading provider of premium network hardware solutions. Our foundation is built on a large-scale operational facility, a robust in-house technical team, and a vast inventory valued at over $10 million.

We are authorized partners for top-tier brands including NVIDIA (Mellanox), Ruckus, Aruba, and Extreme Networks. Our core portfolio encompasses original, brand-new equipment such as high-performance network switch systems, network card (NIC card), wireless access points, controllers, and structured cabling solutions.